Could use a nonparamtric test sugestion...

[nonparametric_princess]

I'm dealing with 12 data points, in the for (Xi,Yi), and I was looking for the optimal nonparametric test to determine if the two data points are correlated or not.

My initial idea was the use of a 2x2 contingency table, and a GOF test to show that they are not independent. I'm weary of this approach because I had to choose such random parameters each bin to have enough observations, and this just seems to round about and weak.

Any Suggestions?? Thanks :-)

[Karabiner]

I am not sure whether I understand your question and your data structure and your
use of the term "nonparametric" correctly, but the usual correlation coefficents for
ordinal and for ranked data would be kendal's tau or Spearman's rho.

Kind regards

K.

[nonparametric_princess]

Thank you so much! I looked into both of these, and Spearman's rho was perfect.